Probabilistic shaping communication system aided by neural network distribution matcher in data center optical network

Zexuan Jing; Qinghua Tian; Xiangjun Xin; Yongjun Wang; Dong Guo; Xia Sheng; Chao Yu

doi:10.1002/mop.32930

Probabilistic shaping communication system aided by neural network distribution matcher in data center optical network

Zexuan Jing, Qinghua Tian^*, Xiangjun Xin, Yongjun Wang, Dong Guo, Xia Sheng, Chao Yu

^*Corresponding author for this work

School of Information and Electronics

Beijing University of Posts and Telecommunications

Research output: Contribution to journal › Article › peer-review

1 Citation (Scopus)

Abstract

A neural network (NN)-assisted probabilistic shaping (PS) distribution matcher is proposed, in which the model is simplified by a structured optimization method. The NN algorithm can encode the information sequence, making the signal obey the Gaussian distribution, and can directly restore the received signal. In addition, the algorithm uses the novel training method at both ends of the transmitter and receiver so that the system performance is significantly improved. PS system verification experiments have been carried out under 16QAM-DMT modulation format. Under the hard decision forward error correction (FEC) threshold of 3.8*10⁻³ BER, the proposed system achieves 1.1 dB improvement compared to the traditional 16QAM-DMT system.

Original language	English
Pages (from-to)	2274-2278
Number of pages	5
Journal	Microwave and Optical Technology Letters
Volume	63
Issue number	9
DOIs	https://doi.org/10.1002/mop.32930
Publication status	Published - Sept 2021

Keywords

machine learning
optical fiber communication
probabilistic shaping

Access to Document

10.1002/mop.32930

Cite this

@article{d634848f291c49ea9f9f14f0afa46e4c,

title = "Probabilistic shaping communication system aided by neural network distribution matcher in data center optical network",

abstract = "A neural network (NN)-assisted probabilistic shaping (PS) distribution matcher is proposed, in which the model is simplified by a structured optimization method. The NN algorithm can encode the information sequence, making the signal obey the Gaussian distribution, and can directly restore the received signal. In addition, the algorithm uses the novel training method at both ends of the transmitter and receiver so that the system performance is significantly improved. PS system verification experiments have been carried out under 16QAM-DMT modulation format. Under the hard decision forward error correction (FEC) threshold of 3.8*10−3 BER, the proposed system achieves 1.1 dB improvement compared to the traditional 16QAM-DMT system.",

keywords = "machine learning, optical fiber communication, probabilistic shaping",

author = "Zexuan Jing and Qinghua Tian and Xiangjun Xin and Yongjun Wang and Dong Guo and Xia Sheng and Chao Yu",

note = "Publisher Copyright: {\textcopyright} 2021 Wiley Periodicals LLC.",

year = "2021",

month = sep,

doi = "10.1002/mop.32930",

language = "English",

volume = "63",

pages = "2274--2278",

journal = "Microwave and Optical Technology Letters",

issn = "0895-2477",

publisher = "John Wiley & Sons Inc.",

number = "9",

}

TY - JOUR

T1 - Probabilistic shaping communication system aided by neural network distribution matcher in data center optical network

AU - Jing, Zexuan

AU - Tian, Qinghua

AU - Xin, Xiangjun

AU - Wang, Yongjun

AU - Guo, Dong

AU - Sheng, Xia

AU - Yu, Chao

PY - 2021/9

Y1 - 2021/9

N2 - A neural network (NN)-assisted probabilistic shaping (PS) distribution matcher is proposed, in which the model is simplified by a structured optimization method. The NN algorithm can encode the information sequence, making the signal obey the Gaussian distribution, and can directly restore the received signal. In addition, the algorithm uses the novel training method at both ends of the transmitter and receiver so that the system performance is significantly improved. PS system verification experiments have been carried out under 16QAM-DMT modulation format. Under the hard decision forward error correction (FEC) threshold of 3.8*10−3 BER, the proposed system achieves 1.1 dB improvement compared to the traditional 16QAM-DMT system.

AB - A neural network (NN)-assisted probabilistic shaping (PS) distribution matcher is proposed, in which the model is simplified by a structured optimization method. The NN algorithm can encode the information sequence, making the signal obey the Gaussian distribution, and can directly restore the received signal. In addition, the algorithm uses the novel training method at both ends of the transmitter and receiver so that the system performance is significantly improved. PS system verification experiments have been carried out under 16QAM-DMT modulation format. Under the hard decision forward error correction (FEC) threshold of 3.8*10−3 BER, the proposed system achieves 1.1 dB improvement compared to the traditional 16QAM-DMT system.

KW - machine learning

KW - optical fiber communication

KW - probabilistic shaping

UR - http://www.scopus.com/inward/record.url?scp=85107435069&partnerID=8YFLogxK

U2 - 10.1002/mop.32930

DO - 10.1002/mop.32930

M3 - Article

AN - SCOPUS:85107435069

SN - 0895-2477

VL - 63

SP - 2274

EP - 2278

JO - Microwave and Optical Technology Letters

JF - Microwave and Optical Technology Letters

IS - 9

ER -

Probabilistic shaping communication system aided by neural network distribution matcher in data center optical network

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this